Also, 2024 still seems aggressive to deploy self-driving for real.
I also don't think they will try for self-driving out of the gate. I know a good number of the experts in this field and no one thinks full autonomy is ready. The moat that Waymo has built with their particular kind of testing and iteration is diametrically opposed to how companies like Tesla, Lyft, Uber etc have tried to do self-driving. Even they are going to be restricted to the easiest of locals for years.
I have to strongly disagree with this. The touchscreen is great for the vast majority of things it replaces. The only physical controls I ever miss are for temperature (which is a very minor complaint) and windshield wipers (though I would also accept auto wipers that actually work). I do think the UI smoothness and touch latency could be slightly improved, and Apple would do better, but Tesla is already so far ahead of every other car manufacturer there that it's not even funny.
If the windows start to fog during one of those situations, taking your attention off the road to look and swipe is very dangerous.
Edit to add: Is the voice processing being handled on board? If not, then if your signal fails....
The Model 3 at least is remarkably quiet; hard rain and street noise and dampened a lot. To the point where we felt it was almost eerie the first time we took it home. Voice recognition has been perfect in all conditions.
You do need to speak loudly, and there's an unfamiliar concentration aspect to using a voice command rather than a switch.
I prefer everything else about the 3.
Both the Tesla and the X5 annoy me when I can't tell the state of the switch by feel. Which is funny, because, IMO, BMW has some of the best-feeling and best-weighted stalks in the business. It sounds silly, but I positively LOVE the way engagement on the steering wheels stalks feel on my X5, and every other BMW I've owned or driven.
Very much agreed on the feeling and weight of the stalks. By far the best in the business. Especially for knowing the difference between a lane change vs. turn signal
I have a Z3M where I can see all the instruments perfectly, but that's just because the steering wheel is fixed and much higher than I'd like.
Here’s an example of various controls I fiddle with while driving: temperature control through a little knob that sticks out; Heated seats through a button with sight curve around borders; Apple CarPlay through a knob.
There’s also no heads up display to present speed and nav onto the windshield.
All these things are good for me to look at the road and not have to precisely touch something with at least a few seconds of focus.
I have friends who love their Tesla and love the touch screen so it may just be my control style that doesn’t work.
I suspect Apple would get this stuff better.
I’m just hoping that they don’t relegate a lot of stuff like climate control to a Siri-only interface.
I think they should prioritise physical controls for stuff you use regularly and I'd be happy to have voice on controls for stuff I use very infrequently.
Like putting the rear window wiper on, barely use it and half the time I accidentality turn it on (or someone else using the car does) and I have no clue how to get it off again. Just saying "siri clean the rear window" or "turn the rear window wiper off" would honestly just be easier.
I wouldn't have minded a HUD that I could turn on and off though.
This is both the most functional and most beautiful design choice I've ever experienced. My favorite part of my Model 3. You quickly come to realize that voice control does everything and that you should never have to reach for a button or fiddle with climate controls, etc.
Tesla realized what all other car manufacturers failed to see, you don't need physical buttons because you don't need buttons of any kind. I'm never buying another antiquated car with physical buttons again.
I think you should consider the tradeoffs here.
Practical safety on the road today, where I don't die because I'm fiddling with buttons, trumps theoretical unmeasurable fearmongering about vague privacy concerns. The reality is that if Tesla were to record every command I ever give the car and post them online against my will, no one will not learn much about me aside how aggressive I want the windshield wipers to be and that I enjoy having a warm butt when I drive.
Talking to electronics does not increase practical safety.
Yes, a car that can slow down to avoid collisions is a form of safety. Seatbelts are another form. A car that does various things to avoid driver distraction is another form of safety. Just because you can think of a more immediate safety system does not make other increases to safety in some way "not practical" or "not real increases".
There are many things, small and large, that go into practical safety of a car
Safety belts work even if I am unconscious, a car that slows down works even if I am unconscious, airbags work even if I am unconscious, so they are 100% safety devices, as in "they keep me safe when everything else failed and things are already out of my control".
Hands-free calls are a safety feature because they are designed to eliminate the distraction of operating a phone.
But they are a workaround, the real safety feature is "operating a phone while driving is prohibited by the law" because the simple act of talking is a distraction.
Voice commands are a workaround because the real safety feature is "touch screens are a distraction and you shouldn't use them while driving"
If we remove touch screens from cars, there is no need for voice commands.
Most of the sources of distraction in modern cars are caused by functions that do not belong to cars and are hard to operate.
I work for a large insurance company based in Europe whose primary business is car insurance (>10 million clients).
We launched a two-year program to monitor the distractions of our customers (we especially monitored the phone's activity, thinking it was a major issue) and the primary sources of distraction while driving are (in reverse order of importance)
- adjusting settings or operating car controls: less than 1%
- eating or drinking: 2-3%
- reaching for things: 2-3%
- talking: 4-5%
- multitasking: 5%
- looking at the GPS display: 10%
- texting and driving: 10%
- using a smartphone: 15%
- not looking at the road for too long: ~50%
there are subcategories for each one of the macro-categories above.
For example:
Adjusting car settings like mirrors, seats, the heater is basically risk-free, more than half of the distractions in this category are caused by adjusting the audio volume. It makes sense, muscle memory for settings helps us to not think too much when doing it. Regulating the audio volume usually follows some unexpected event that forces us to change it. Still only 1/200 distractions are caused by it.
Eating or drinking is mostly eating here, I guess in the USA it would be the opposite. Bringing hot beverages inside the car is quite common there, here is quite rare. People spill the coke on their shirts or choke on the water anyway. It is also very common to frequently stop at motorway restaurants here to eat, drink, use the toilets, rest a little bit and eventually refuel. It helps to keep this kind of distractions to a minimum.
Reaching for things is less dangerous when the driver picks up things from the driver's or the passenger's seat, opening the glove compartment is two times as dangerous, picking up something from the pedals area is the most dangerous of them all.
Talking is interesting because there's no difference between people talking with other occupants and people talking on the phone. What's less interesting and probably obvious is that 8/10 happen when people are arguing. But 2/10 are people talking normally like they would to a car to activate the wiper.
Multitasking is when people do other things unrelated to driving the car itself. For example, handling pets is more dangerous than applying makeup that is more dangerous than handling children. The good news is when people carry kids around they drive more carefully.
The GPS display is a major source of distraction, no matter what voice people chose to give them directions, the tendency to not trust other people's instructions makes people look at the screen to confirm what they heard. Also, there are times, not rare, when the instructions are not clear or haven't been updated to the actual road conditions (a roadblock for example.
Despite the technological improvements, texting and driving is still very popular, especially among younger generations. More importantly, we counted using chat apps in this category and it's not surprisingly the vast majority of the cases.
Using a smartphone is an interesting one: while the "phone" part of the smartphone has been standardized and is easy to remote nowadays, most of the time people use the "smart" part of the smartphone, which cannot be remoted. They scroll the Instagram feed, they take pictures, make videos, and, despite having the freehands at their disposal, people prefer to send voice messages. Which involves removing your hands from the wheel.
Did you notice how many times talking and/or hearing voices caused a distraction?
But the most important of them all is not looking at the road. It sounds obvious once you hear it, but it happens a lot more than people think, 100 times more than turning the volume up or down. It could be an external event like an accident, an emergency vehicle passing by, a fire, or something that attracted the driver's attention like a billboard or a beautiful person walking on the sidewalk - it happens more often than you can imagine -. These are the "good" ones, often they don't last enough to cause severe accidents and people just end up rear-ending each other at the traffic lights. The real killer is being lost in one's thoughts, zoning out, keep driving on autopilot while you are not there anymore, and don't realize you are still in a car moving at 50km/h.
That's how accidents happen most of the time.
That's why I think that cars that slow down automatically or correct your errors are a breakthrough, they already exist on the market, have no voice control but they can alert you with a warning sound if they detect that you are distracted behind the wheel.
A car that adjusts the temperature of the heather using voice commands is not equally useful, I would argue that is completely useless unless fixing the temperature in the car will become a dangerous activity in the future, but honestly, I don't see that happening very soon.
So yes, small things are important, but BIG things are even more important.
Let's fix those before thinking that cars need UI/UX designers more than safety engineers.
Before "disrupting" the car industry for the sake of changing things and adding the cool factor to something that doesn't really need it.
p.s.: My opinion is based on the experience gathered doing my job, I'm not against voice commands because I don't like them, but because they can be a source of distraction per sé.
p.p.s.: it took me time to put this together, so please if you don't like it don't downvote it without leaving a comment. Thanks.
Totally agree.
> Let's fix those before thinking that cars need UI/UX designers more than safety engineers.
"Let's fix big things before we even think about fixing small things". Which is silly. We can try to improve multiple things at once. We can improve easy small hanging fruit before we improve harder things. This is a false dichotomy I think.
Anyway, all tesla vehicles already have automatic emergency braking, so I don't really understand why you went on such a long rant to argue for adding that instead of voice commands. The car under discussion has both, which seems better than having just one or the other.
Which is naif.
There is a limited amount of time and resources that a company can spend on features , prioritizing big things is always the sensible choice.
In the specific case of cars, they can't be patched after reaching the market and usually a recall is issued only after something serious has already happened.
Thake the example linked below that made the front page: A Tesla car speeds on autopilot when both seats are reclined in sleep mode.
Tesla overlooked safety to provide useless features.
If they had put big things first, the car would completely stop if the driver seat is being used as a bunk bed.
And why the autopilot can go over the limits?
They didn't think about it, because the amount of time engineers can spend debugging issues and QA finding problems is limited.
(someone in the comment wrote «Teslas often “shadow brake” when driving under overpasses or overhead signs because the radar system gets confused.»)
The way echo speakers are implemented is that the device only determines if you used the wake word and then sends it to the cloud service. However this wake word detection is sloppy and turns on more often than it should. It's filtered out on the cloud service side, but it creates a situation where extremely private moments are uploaded. For end users this means that you can't be certain when it's recording vs when it isn't....
IDK how it's implemented in cars, but the STT features I've seen require the explicit press of a button and IIRC it's inferenced locally on the car's hardware. Which is possible thanks to the car's high energy usage and price. But it might depend on implementation and it might send telemetry back to the manufacturer.
I shouldn't say to my car what I want the car to do, I should do it.
Also, I listen to loud music when I drive.
Imagine you are driving and the kids start activating car functions because it's fun.
Thanks but no thanks.
When you press it, music pauses/dims, so loud music isn't going to get in the way of you being heard. Kids can't just yell "car, do X" because they can't hit the button on the steering wheel.
There are valid reasons to dislike voice-controls, but the ones you've picked are not issues with the specific system in question.
Similarly in many cases with windshield wipers, headlights, windows, temperature controls, etc.
Drive by wire systems are simply not completely mechanical anymore, but you don't have to ask them to do something for you, they translate a mechanical input into an electric signal and then in a mechanical output.
There is no understanding involved.
I hope you realize that.
Th gas pedal is still a moderately precise mechanical actuator that can be easily and accurately measured regardless of who uses it, you can also use a stick or a brick to accelerate, the car doesn't care as long as you move the pedal.
You can also be a mute and use it.
big difference uh?
But they also have an extensive partnership with Waymo: https://self-driving.lyft.com/riders/
I also know that both Alphabet and GM have large stakes in Lyft. I bought stock in Lyft with the expectation that they will one day be acquired by one of their partners.
If you really want to make this fun, you don't even need to know as you can use traffic aware cruise control, a feature many cars have, to set and forget your speed on any road.
I have two issues with the UI/Human interface that Telsa has followed. First, their automatic wiper solution is dreadful and while you can instant a wipe action by pressing the button on the end of the stalk I believe the variable rate should be available on the stalk.
The second is that I think the steering wheel would be better served with buttons dedicated to cell phone use, namely accept, ignore, mute, and hangup.
However it should be noted the cars voice controls respond very well and you can easily change the temperature should you want that way, there are a lot of settings easily changed by voice but Tesla does not alert you to them.
The best way to grade their UI though is that I can muscle memory activate many features including one level down options. Of course you can just let the car drive and play with the screen all you want, something I have done to prove a point but the simple fact is...
I rarely if ever, as in nearly never, need to access anything through the screen as its all set and forget. Now I get into other cars and think, damn this has enough buttons and screens to be an airliner.
It won't be but a generation before we look back and wonder how anyone though six gauges, forty buttons, and three or more levers, were ease of operation, Not all cars will be as simple as a Tesla but the age of forty buttons is going away.
(Seriously, count how many buttons are in your car - then try to remember how often you use them)
(Photo of someone else's dash: https://cars.usnews.com/static/images/Auto/izmo/323783/2010_... )
It also totally sounds like something Apple would do.
[0] ATP 409 https://atp.fm/409
But they have to be used appropriately.
A touchscreen for wiper control, or to turn on defrost, or to flash your headlights at another car is NOT appropriate.
A touchscreen for a nav display, or to set infrequently used preferences is very appropriate.
I test drove a model 3 and ... couldn't buy one because of the dash.
I think there should be - always - status in line of sight of the driver.
The model 3 dashboard in the center of the car is just a bunch of bathwater with no baby.
Some people won't care. I've had friends say they got used to it.
I can... almost... think of it like an off-road motorcycle with a handlebar and no instruments.
But personally I couldn't buy in.
Related, I think the controls on the model 3 have also been over-minimized. The scroll wheels have the feel of a $3 computer mouse, and autopilot overloaded on the gear selector is cheap, not elegant.
Nobody has figured it out out yet, and a 'moat' for a market that is basically impossible to a big money-maker is not a 'moat' at all.
The quest for self driving is not the same as the quest for the first commercial deployment. Solving general self-driving is the real quest.
And I don't trust any experts to predict this correctly, as nobody knows yet what the solution actually is. Some experts say one thing, others say other things.
I think this might just come down to personal preference. After owning a Model 3 for a while now, I can't stand my Audi (or other cars) that aren't touch screens. There are far too many buttons and dials. It is overwhelming (and ugly to me). The non-touch UIs are clunky and confusing and often far more distracting for me as I try to figure them out.
There's room for more than one car company like this. If Apple brings high quality, tight fit and finish, elegant design, software focus, vertical integration (including their new M1 processor tech), logistics support (charging station network, maintenance, etc), and "next-level batteries" to the mix, then I'm sure they'll find space in the market.
If Apple can execute with electric cars as they've executed with consumer electronics, then they'll be competing not just with Tesla, but directly with Audi, BMW, Mercedes, Lexus, etc.
Honestly, if Apple bought out the existing non-Tesla charging networks and dedicated even a sliver of their UX talent to improving the charging experience, they would really push things forward. This is, of course, assuming that buying the charging networks would make it past the anti-trust people.
Would I have to flip my iCar over and plug into the bottom center?
I don't know if Apple's quality is going to translate to cars either unless they poach a bunch of people. Safety features will compromise the quality at some point, but maybe they're good enough to work around it elegantly.
Plus people paid a premium for Apple's quality, but are they willing to do so when that premium is multiple thousands of dollars extra for a car? They could afford low/zero cost financing if they wanted to.
Tesla is still far from that. I've yet to meet a tesla owner who's car just "Works" without any issues or fiddling.
GM's & Nissan's EVs fit that bucket. But they are designed as "buget" cars and dont evoke a desire from buyers to be excited about them on a day-to-day basis.
My interactions with Tesla service since acquiring my Model 3 in December 2018:
(1) Install rear spoiler (out of stock when car originally ordered - free service)
(2) Upgrade self-driving computer to version 3 (free service given the car had the FSD option)
(3) Rotate tires
All three were performed at my home, at my convenience, by a tech who just showed up and did his business without needing my attention.
Car and experience have been great - so good that my wife leased a Model 3 in 2019 and will be replacing it with a Model Y when the lease expires.
In the car market, being a different brand is a differentiator. Ferrari, McLaren, Bugatti, Lamborghini, and Porsche don't offer vastly different products. But the car market is heavily driven by people wanting something different from their friends. Once Teslas become popular enough, people will want something else just to be different.
Extremely good build quality? Accessories notwithstanding, most of their products look and feel like they were built extremely well.
Even their failures like the butterfly keyboard looked and felt like they were built to extremely tight tolerances.
To be fair to the designers, the height budget they were working with is an extremely hard target to hit for a part that gets as much physical abuse as a keyboard. I was pretty impressed with how passable the keyboard was for typing on, given the height constraint. (However, I was not impressed with the dropped actuations, double actuations, inconsistent key feel after a few months use, broken keycap that my 1-year-old keeps removing and hiding around the house, etc. Overall it ended up being a catastrophically bad design based on poor reliability.)
The biggest problem at company scale was that they couldn't revert to something like their previous keyboard in a shorter time frame. The lesson is to avoid completely locking the whole rest of the system design around a brand new untested part for which there is no alternative. It would have been better to test the new keyboard design in a context like an external iPad keyboard where if it turned out to be a dud there would be an obvious path to an alternative, and customers wouldn't be left with a broken $2000 machine.
I could believe that their test rig didn't adequately simulate the motions/stresses of finger keypresses.
Uh, the butterfly isn't used in Apple keyboards anymore. They reverted to the old scissor switch design some time ago.
It would have been much better if they had figured out how to cut their losses and revert the keyboards in 2016 or early 2017, even though that would have taken a substantial redesign of the rest of the laptop internals.
Selling 4.5 years of laptops with keyboards that broke easily under ordinary use was a huge black eye for the company.
Given that Apple is still selling, or has only just stopped selling models with the latest butterfly keyboard - the issue may have solution, but hasn't gone away. Millions of laptops have faulty keyboards which, when they break, will be replaced with another faulty keyboard. That's an unresolved issue.
Instruct an engineer to design something that is not crap and they will do it, and it will be mass produced with tight tolerance if specified like that. On the manufacturing side... Apple does not manufacture anyway.
And that brings me to: a car is not exactly the same thing as your small household appliance. Like we saw with Tesla, it is not trivial to do with a good build quality, both on the design front and on the build front (way more difficult than small and mostly static objects) -- and it is harder to reject a car with a not stellar build quality because that's more expensive and a huge part of the value is in the mechanics. And quite related: for now Apple does not manufacture, if they start to do cars, will they? Otherwise who will?
I bet they support CarPlay, at least :). I have CarPlay on all my cars excdept for my Tesla, and it still annoys me. I won't buy another Tesla that doesn't support it.
It'll only work with an iPhone.
Really? Why are you not curious what the differentiator of all the car companies out there is? Why don't we just have one car company?
(Hint: There is so much demand, so many different tastes, budgets and features and regulation, that a single company will never serve all of them. Tesla is a niche product and I for one would never buy one)
The iPhone was special in the regard that there simply was NO competition. We are not talking EV vs combustion (which was Tesla's main step, and even still, everyone knew EVs would work and how they work, but nobody was ready to fully commit to them yet). We are talking horses vs. cars. Apple managed to establish themselves in a market that most companies didn't even think exist...
So no comparison here. All Tesla has done is make EVs more approachable and "cool" for the masses. However they have literally ZERO differentiator compared to established car companies. They are going to tank so hard stock wise in the next 10 years that investors will wonder what hit them.
> they have literally ZERO differentiator compared to established car companies. They are going to tank so hard stock wise
Their stock is insanely overvalued, no argument there. But they have one of the largest installed and fastest growing DC charging networks, software that doesn't suck, and the most capable and fastest improving driver assist system enabled by a proprietary custom ASIC. That's far from "ZERO" differentiators, especially in a world where other manufacturers keep faceplanting on their software efforts over and over again.